This invention relates generally to techniques for improving performance in communication systems using code division multiple access (CDMA) and, more particularly, to techniques for reducing interference among multiple simultaneous users of communication systems of this type. CDMA is a spread-spectrum form of communication, by which is meant that the information is transmitted as a signal that occupies a bandwidth in excess of the minimum necessary to send the information. In CDMA, spectrum spreading is accomplished by means of a code that is independent of the data being transmitted. The data modulates the code prior to transmission and then a synchronized code generator at a receiver despreads the received signal and recovers the data.
Multiple users of a CDMA system share the same frequency band and transmit at the same time, but are distinguishable because each user is associated with a different code and the codes are said to be orthogonal, i.e., in theory they are independently detectable in the receiver. CDMA has some significant advantages over alternative techniques for wireless digital communication. One disadvantage is the inherent presence of multiple access interference (MAI) when multiple users are transmitting simultaneously to the same receiver. In a practical communication system, perfect orthogonality is lost because, even if perfectly orthogonal codes are used for each user signal, signal fading and frequency errors result in loss of orthogonality. Further, since there is only a limited number of orthogonal codes, pseudo noise (PN) codes are also used in CDMA systems. Because PN codes are not perfectly orthogonal, the user signals will, to some degree, be subject to mutual interference, i.e., MAI. In the context of a mobile communication system, a single base station receiver may be simultaneously receiving signals from multiple users in the same geographic xe2x80x9ccellxe2x80x9d served by the base station. An important measure of performance in communication systems generally is signal-to-noise ratio (SNR). More specifically, in digital communication the ratio is termed Eb/N0, or energy per bit noise power density. In this specification, the term SNR will be used, but it will be understood that this refers to the more specific term Eb/N0 
In CDMA, the noise component in the signal-to-noise ratio includes not only random, and often uncontrollable, noise introduced in the radio-frequency (RF) communication channel, but also any interference introduced by transmissions of other users. For any particular user, user #1, for example, the transmissions, if any, of user #2 and users #3 through #N are noise contributions so far as user #1 is concerned. In early versions of CDMA, multiple access interference (MAI) was recognized as a natural consequence of serving multiple users in a single cell, but the only solution considered was to reduce the number of users that could be served by a single receiver. The presence of MAI affects system performance by limiting system capacity, i.e., the number of users who can share the transmission bandwidth at the same time, and degrading voice quality for each user.
More recently implemented CDMA systems have attempted to mitigate MAI in order to achieve the twin benefits of increased capacity and improved voice quality. Prior to the present invention, MAI mitigation has been achieved to some extent by means of a technique referred to as multistage parallel interference cancellation (MPIC). Initially, MPIC detects signals for all users and then estimates the multiple access interference imposed on each user by combining the detected signals of the other users. Interference cancellation is performed by subtracting from each user""s received signal the estimated MAI for that user. In a second stage of cancellation, the corrected signals are used to compute second-stage MAI cancellation for each user signal. Although the MPIC approach can provide a significant improvement in system performance, one serious shortcoming is that the results may diverge instead of converging on a progressively more accurate MAI value. The technique relies on accurate detection of data bits for all users at each iteration. If the bit error rate (BER), which is a measure of the effect of noise, does not improve at each iteration, especially in the first few iterations when the interference levels and the BER are higher, then the technique is often prone to error.
More specifically, every stage of MPIC cancels MAI as estimated from the detected data bits of the previous stage. Ideally, performance is improved by each successive MPIC stage, since more and more interference is canceled. Unfortunately, errors in the MAI estimates are inevitable because of non-zero bit error rate (BER), and performance improvement with additional stages is not guaranteed. Recently, partial cancellation of MAI was proposed, where the partiality is implemented by weighting the MAI in proportion to confidence in each bit estimate. If a more reliable MAI estimate is used, MPIC is more likely to converge. (See D. Divsalar and M. K. Simon, xe2x80x9cImproved CDMA Performance Using Parallel Interference Cancellation,xe2x80x9d IEEE Transactions on Communications, vol. 46, no. Feb. 2, 1998.) Others have proposed an adaptive MPIC, in which weights are optimized through minimizing the Euclidean distance between the received signal and the weighted sum of each estimated signal by LMS (least mean squares) algorithm. (See. G. Xue, J. Weng, T. Le-Ngoc and S. Tahar, xe2x80x9cAdaptive Multistage Parallel Interference Cancellation for CDMA,xe2x80x9d ICC""1999, Vancouver, Canada.)
It will be appreciated from the foregoing that prior art approaches, and specifically MPIC, for reducing the effect of MAI are not completely reliable. Even the enhanced versions of MPIC may not always converge if the BER is significantly high. A simple and more reliable alternative would be beneficial in many applications of CDMA, and the present invention is directed to this end.
The present invention resides in a method and corresponding apparatus for cancellation of multiple access interference (MAI) in a code division multiple access (CDMA) communication system. To overcome the shortcomings of multiple parallel interference cancellation (MPIC) described above, the present invention uses space diversity gain to achieve a high signal-to-noise ratio (SNR) for each received user signal, thereby obtaining a low initial bit error rate (BER). Successful multiple access interference (MAI) cancellation depends on having a low BER at every stage. A low initial BER results in more effective MAI cancellation in the first and subsequent stages of multistage parallel interference cancellation (MPIC). As described in more detail later in this specification, a low BER is obtained as a direct result of the high SNR achieved from the use of space diversity gain.
Briefly, the method of the invention comprises the steps of: simultaneously receiving at spatially separated multiple antennas, signals transmitted from multiple users, the signals from at least some of the users following multiple propagation paths and arriving at the antennas as resolvable multipath components; for each user signal, selecting from among all of the multipath components received at all of the antennas a subset of resolvable multipath components of the user signal that provides the highest signal-to-noise ratio; and coherently combining the selected subset of multipath components to obtain for each user a single combined received signal for which signal-to-noise ratio has been maximized by use of space and time diversity. It will be understood that, in the context of the present invention, the term xe2x80x9csubsetxe2x80x9d may indude the complete set. Thus, a subset of n multipath signal components selected from among L multipath components derived from each of M antennas, may include as many as Mxc3x97L signals. Therefore, the single composite signal for a particular user may be derived from as many as Mxc3x97L signals. The method further includes the steps of estimating multiple access interference (MAI) for each user; and subtracting MAI from the received signals to obtain estimates of the user signals compensated for MAI. Preferably, the steps of generating estimates of MAI and subtracting them from the received signals are performed by one or more stages of parallel interference cancellation.
More specifically, the method of the invention includes the steps of receiving signals from multiple users of a CDMA communication system simultaneously, at a plurality of spatially separated antennas, wherein at least some of the signals reach the antennas as multipath components traversing a plurality of different propagation paths; converting the received signals to digital form; separating the signals received at each antenna by multipath component and by user; for each user, selecting a number of signals from all of the multipath components derived from all of the antennas, wherein the signals are selected as having relatively high signal-to-noise ratio; coherently combining the selected signals to produce a high signal-to-noise-ratio (SNR) signal for the desired user. Finally, the method includes performing multistage parallel interference cancellation (MPIC) on the user combined signals obtained by the preceding steps. The method employs both time and space diversity to provide more accurate estimates of user signals. Therefore, multiple access interference is canceled more reliably than by MPIC alone.
Preferably, the step of performing multistage parallel interference cancellation includes spreading the separate composite user signals, using spreading codes associated with the users; computing a set of estimated MAI cancellations, obtained as partial sums of the spread signals, wherein the partial sum for each user is computed as the sum of all of the other user""s signals; subtracting each partial sum from the received but not yet despread signals, to obtain a first estimate of spread user signals compensated for MAI; then despreading the spread user signals compensated for MAI in the preceding step; and repeating the preceding steps at least once more to obtain more accurate estimates of the user signals compensated for MAI.
The invention may also be defined in terms of receiver apparatus for cancellation of multiple access interference (MAI) in a code division multiple access (CDMA) communication system. The apparatus comprises multiple antennas for receiving signals from multiple users transmitting simultaneously to the receiver apparatus, the signals from at least some of the users being received as multipath components traversing direct and indirect paths; multiple radio-frequency (RF) receivers coupled to respective antennas; signal processing means coupled to each antenna, for separating the received signals by user and by multipath component; means for selecting for each user a subset of signals from all of the multipath components received through all of the antennas, wherein the selection is made on the basis of maximizing signal-to-noise ratio; means for coherently combining the selected signals for each user, to obtain an initial set of estimated user signals with high signal-to-noise ratios; and a multistage parallel interference cancellation (MPIC) circuit, connected in a cascade arrangement with the foregoing apparatus elements, to provide cancellation of multiple access interference cancellation that operates to reduce the bit error rate for each user. The initial set of estimated user signals derived by the apparatus provides more efficient reliable operation of the MPIC circuit.
In coherently combining the selected signal components for each user, timing and tracking information for any multipath component received at multiple antennas can be shared to facilitate processing.
It will be appreciated from the foregoing summary that the present invention represents a significant advance in CDMA technology. In particular, the invention provides a more reliable approach to cancellation of multiple access interference, and thereby increasing system capacity and improving transmission voice quality. Other aspects and advantages of the invention will become apparent from the following more detailed description, taken in conjunction with the accompanying drawings.